***OFFICIAL Coursera Game Theory Spring 2013 Thread***

A small number of us (Celtic, Rilla, and I) are part of Coursera's Game Theory class with the link here: https://class.coursera.org/gametheor...02/class/index.

We thought it'd be a cool idea to study together in FTR in a public thread and see how it works. It's a good start that can lead to an awesome commune dynamic in the future (members of a forum studying together and perhaps wufwugy even learning a thing or two). If you're interested in joining, just click the link.

Coursera offers certificates to those who've completed the course which is pretty cool. Any other questions about this course or Coursera can be asked here.

I've no questions yet to pose to Celtic or Rilla. FWIW Celtic, I haven't started week 2.

GAME ON!

Enrolled. Ship it.

Oh man, how do i find time to do this :/

Inb4 spoon laughs at us

i enrolled. i dont know much about game theory but i think ill try problem set #1

We are given two attempts for problem set 1 but only one attempt for future problem sets.

This looks like fun, I'm gonna enscribe once I get home.

I've watched the first 2 videos. The 2nd one sent me to sleep, so I'll have to try again.
I've also downloaded the Nash movie. Looks quite good. Watched first half hour.

Great idea BooG, posted this on Facebook too.

I'm planning on doing the Harvard theory on Justice in Spring so I'm out sorry :(

Quote:

---

Originally Posted by Pascal

Great idea BooG, posted this on Facebook too.

I'm planning on doing the Harvard theory on Justice in Spring so I'm out sorry :(

---

All credit belongs to rilla for the idea. I just created the thread.

BTW if you don't think the Facebook thing is a good idea LMK and I can take it down.

I doubt any freetrollers will come in. Let's see how it works.

The first problem set is due Sunday. I took a look at it and it's gibberish through eyes that know nothing about game theory. I'll be around FTR Sunday to take in the sights of this class and hammer out the answers. I don't think it'll be a bad idea to straight up post answers in the thread. People who want to laze through the class may, others who fall behind a bit but don't want to be discouraged in later weeks may as well take them to their benefit. And for anyone who genuinely just needs to see an answer to vet their own work, it's invaluable.

I'm beholden to other business til Sunday though.

Post the problem set, I'm curious.

Meh, I'd argue that it's a pretty bad idea to post answers to problem sets here. That isn't the purpose of this thread. We're given the answers after we attempt the problem set from here on in (problem sets 2 and above). It leads to people joining this thread for the wrong reasons.

I don't want to be seeing answers her until after Ive submitted mine, so after the submission date please.

Ive now watched all week one videos, Game theory is new to me, Im just about grasping it. The formulas that are being used are a little alien to me, but im not letting that bother me.

Having the answers before you submit as a sanity check or verification helps you solidify your thinking.

If their answer differs from yours, you re-vet to find your error or theirs. If their answer matches, you're comfortable with the work you've done.

Having the answers after you've submitted only serves to highlight those sorts of problems as things you've gotten wrong.

I'm not asking for your answers, just saying I'm thinking about posting mine when I get em. In spoiler tags.

Sigh was gonna post suggesting post them in spoiler tags if you wanna post them but Rilla got them first :( Other option would be to just post and say you've done them and you'd like to compare answers and people can PM answers.

How's the course looking so far to you celtic?

quiz #1 question #3, i guessed the fish answer and im not really satisfied with their explanation. guess ill finally fire up the video

i finally got all the problem/quiz stuff down and rather enjoyed it. good idea guys. lets all get smarter

Quote:

---

Originally Posted by a500lbgorilla

Having the answers before you submit as a sanity check or verification helps you solidify your thinking.

If their answer differs from yours, you re-vet to find your error or theirs. If their answer matches, you're comfortable with the work you've done.

Having the answers after you've submitted only serves to highlight those sorts of problems as things you've gotten wrong.

I'm not asking for your answers, just saying I'm thinking about posting mine when I get em. In spoiler tags.

---

It's a slippery slope but you're obviously going to do what you will. I guess we're mature adults though and what you do with the answers you establish or receive is your business.

Can I make a request though? Can we keep the question discussion and final answer completely separate? Perhaps two different spoiler tags? I know I'm pretty nitty/nerdy about shit like this.

i think its easy enough to explain the answer you know without explicitly giving the answers, so lets just do that when helping eachother. its kinda nitty but i do keep clicking the honor code checkbox =p thats a good trick

Quote:

---

Originally Posted by Pascal

How's the course looking so far to you celtic?

---

Ok, The algebraic notation , im not familiar with. But I can grasp the examples in the video tutorials and arrive at the correct answer ( some of the time)

Ive not even looked at the quiz questions that have to be answered by the 20th.

I dont feel ready to tackle the questions yet and ive not looked. When I do feel ready, I wont be looking in this thread, in case someone posts their answers. The temptation to peek at the spoiler will be too great.

I would rather submit my own wrong answers and learn that way, than tweak my answers to match some that look more correct.

You're all nits. Even if we phase-shifted this into a cheating ring, the worst result would be we don't learn.

And what's the point of taking the class if you don't learn?

But it's immaterial now. See if we can't figure out something better on Sunday as it seems I'm the only one even threatening to post answers.

Seems nobody wants your damn answers rilla. :D

Celtic, if you need help with the notation, I can hop on Skype tonight (Korean time) and review them quickly. I assume they're gonna be pretty important and they're not too hard to learn. Or, you know, Google.

Appreciate the offer. My Skype is off till i get my new Note2 (friday, i think) . Ill give it a google .

Well crap, this sounds fun and awesome and I feel left out. I guess my options are to spread myself way too thin by signing up, to drop my Justice course on edX and take this one instead, or to wait until next semester and take the class with . . . . . Pascal *vomits all over myself*.

http://25.media.tumblr.com/tumblr_mb...qizno1_500.gif

:(

First attempt

6.58 / 9.00

I take offense to a small part of 1-6 "1600 people had named 50 and obviously they hadn't gone through all the reasoning..." where they pose the Keynes Beauty Contest question. They say pencil to paper the answer is 1. And pencil to paper the answer is 1. But when you put the question to the real world, the answer isn't 1. They show this and should expect this. People don't have pencils or papers or recognize the application of Nash's equilibrium to the problem before them.

So when they wonder why so many people guessed 50 that tells me that they're failing to calibrate their thinky-thoughts to reality. It's not the 50-guessers that didn't run through all the reasoning, it's the instructors who didn't by expecting everyone to hammer in 1.

When you ask people to guess a number 1-10, people say 7, people commonly know this or will answer to themselves 7. It extends that when you ask for a number 1-100, the number will be 77. Undercutting that as we so cleverly would puts us at 50. Making 50 a good guess to give.

To undercut 50 and go with 33 requires us to have an expectation that the majority of people will be sitting at 50. This is an expectation we can't have without having already calibrated ourselves to the reality of the people around us.

To undercut 33 to 22 because you assume everyone was confident enough to undercut 50, is again floating on reasoning that needs to be calibrated to reality.

Occams Razor translates into a lazy razor for people. The easiest answer to give is the first one that pops in there. The slightly more difficult one to give is going one step further to your advantage. And this continues, but not til infinity, since no one has time or energy for that. Unless they're skipping ahead as their first step. People will more probably laze in their answers over investing effort in their answers. And they'll always be a scatter, because we're all god damn snowflakes.

Seems to me, we're in trouble every time we assume everyone is rational. Because even the rational ones are being irrational for holding that very assumption.

If you played this game with the same group of people hundreds of times, though. Then we'll approach 1 being a pragmatic answer.

So we're in the prisoner's dilemma

-1,1 -4,0
0,-4 -3,-3

Why do we say that if we consider the other guy to have chosen whatever, that we either get A or B? Instead of saying we either get A and give A' or get B and give B'.

This subject material is sweet.

edit I'm betting it has to do with how we get those numbers. That if I gave a damn about the other guy, the numbers would reflect that.

which poses the next question.

Quote:

---

Originally Posted by a500lbgorilla

If you played this game with the same group of people hundreds of times, though. Then we'll approach 1 being a pragmatic answer.

---

I did arrive at 1 myself. Simulating the contest with pen and paper and imaginary people.

Quote:

---

Originally Posted by celtic123

I did arrive at 1 myself. Simulating the contest with pen and paper and imaginary people.

---

Yeah, and you should arrive at 1 but those people that you simulated weren't real people. Real people won't follow pen and paper until you teach them how to use pen and paper.

edit and even then, the snowflakes!

"I encourage you to pause the video at this point"

Everytime I pause a video, I have to reload and start at the beginning!

No pain, no gain.

Ive started beating people at rock scissors paper lately.

Quote:

---

Originally Posted by celtic123

First attempt

6.58 / 9.00

---

questions 1 2 and 3 you probably got right

question 4 is derived from knowing the answer to 1, and knowing what it means to be a strictly dominant strategy

question 5 again you have know to what a strictly dominant strategy is. a SDS must beat all the other strategies all the time

#6 i know the right answer but would love for someone to explain it

#7
if p=0 then they would lose money, so setting a high p would be better than 0 (because 0 income is better than negative)

also we know they have to have the same strategy, and we know a pure nash equilibr exists

#8 gotta know what a weakly dominant strategy is

#9 what strategy(s) cant be exploited

the TA posted this about #6 "You assume that your opponent will make the same choice that they did last time. Assuming your opponent doesn't change their mind, can you do any better if you change your own decision? If you can't--even if all your choices are equally bad given your opponent's current choice--then you're at Nash."

Got 'em but damn this is tough.

Am I too late to join?

No just sign up. I signed up late too.

slashthreadgate

Welcome, fronds.

Does anyone understand what's going on in lecture 2-4?

Quote:

---

Originally Posted by BooG690

Does anyone understand what's going on in lecture 2-4?

---

Pretty vague because it is pure theory, but what I gathered from it was, in the first slide they explain that they want to make every problem linear, because that can be solved. And if needed, use heuristics (experienced-based shortcuts) to make it linear. On the second slide about PPAD, it's just a vague explanation where they try to quantify the complexity of certain problems. (complexity as in, how hard it is to solve, how much time/how many computations it will take)

All in all not that relevant to what follows in lecture 2.

What did everyone get on problem sets 2 and 3? i got 4/5 on each. Shoulda been 5/5 on the 3rd one, I was just too hasty.

Quote:

---

Originally Posted by a500lbgorilla

If you played this game with the same group of people hundreds of times, though. Then we'll approach 1 being a pragmatic answer.

---

You don't need hundreds. In the second try with the same group of people the mean was already down to 4. Won't take long to settle on 1.

Yeah, I saw that right after I made the comment. Though they had a heavily reduced number of people in the second one.

I'll probably try to get started on this stuff on Saturday morning, so I have Sunday for getting over any walls. Wish I got into this stuff earlier but people drop like flies in that werewolf game, I'll tell you what.

Got 5/5 on PS2. Will do 3 tomorrow or Monday.

Quote:

---

Originally Posted by BooG690

Does anyone understand what's going on in lecture 2-4?

---

"Look at all the shit us Game Theorists have done."

woot woot, 5 by 5. Really only two new tricks to roll over across all 5 problems but they're basically as straightforward as you think.

How many people did we lose through 2 weeks?

Who's still at it? Me and boog and jv?

Is it too late to join this?

Well, we're two assignments in. You can still do them for half credit, but this week you'll have to do 3 weeks worth of work which does take its fair share of time.

But yeah, I think sign ups are still open.

It told me I got me a 100% credit penalty on the first one (having signed up so late), you can get half credit too?

I dunno where I pulled that half-credit thing from, but that was the impression I was under. Or partial credit, really.

Guys, I'm more into programming, so I have like CS50x and Logic on my plate etc. To those who are doing it, is Game Theory as taught somehow related? Even though I'm quite late, I could jump een eet anyways.

{edit}God damn, hours per week are somewhat limited, so this will be a stretch.

Well it's been definitions and how to calculate mixed nash equilibria. Which is basically making the EVs of each of your opponents choices equal to one another as he does the same to you.

Can't help but observe that we'd probably have more activity in this thread if we had posted the answers.

If only because answer discussion is basically the entire class.

We can easily discuss the last two weeks though? It seems nobody has had trouble since it has been pretty easy until now. Celtic had trouble with algebraic expression but that's about it.

So I had a stray thought.

Beyonce was caught lipsyncing at the inauguration. Beyonce is going to perform at the superbowl.

Can you describe her choice to lipsync or not as a game in the spirit of the class so far?

Bamp

Doing this now. Everyone else keeping up?

This week looks meaty.

I missed part of 2 and all of 3 for the same reason. I thought you had to "get" to the scenarios to actually consider them.

So for instance, I thought all choices to 'not play' were equilibria since all other choices are moot. But one of them wasn't even though it's not essentially part of the tree any more.

And for the other, I thought you assumed you were getting to the subgame in the first place in order to then find its equilibrium.

So the lesson is, even stuff that isn't going to possibly occur still needs to be considered for analysis.

I just finished three and I am getting to four in a few minutes (hours).

are you guys finding much of the material being covered directly relevant to poker?

I came here to comment on that. Week 4 is finally a bit relevant to poker since we're doing extensive form games (games moved in sequence [player 1 then player 2 then player 1, etc.). The ones I'm up to are perfect information though imperfect information games will be covered.

There's a course that will be provided by Coursera which I'm enrolled for which will probably cover games like this more thoroughly: https://www.coursera.org/course/cgt

Edit: Looking at future lectures for this week (which I'm doing last minute), they have one devoted to poker. I've uploaded it for you guys to watch, if you're interested: https://dl.dropbox.com/u/13302835/4%...0%28336%29.mp4

Edit Edit: The course I've linked here hasn't much to do with poker since it's games WITHOUT chance.

The Centipede game is awesome.

Quote:

---

Originally Posted by a500lbgorilla

I missed part of 2 and all of 3 for the same reason. I thought you had to "get" to the scenarios to actually consider them.

So for instance, I thought all choices to 'not play' were equilibria since all other choices are moot. But one of them wasn't even though it's not essentially part of the tree any more.

And for the other, I thought you assumed you were getting to the subgame in the first place in order to then find its equilibrium.

So the lesson is, even stuff that isn't going to possibly occur still needs to be considered for analysis.

---

There's a lecture where he explicitly says this isn't true.

Shit. Answers in spoiler for problem set 4.

Spoiler: Show

Offer every other pirate some gold. Fuck. Stupid me. Fuck.

I WANTED to say 98 because it made logical sense...fuck.

Also, I rushed in writing normal game tables for two of these extended games. Totally fucked up my score. Take your time in writing the payoff tables.

Quote:

---

Originally Posted by BooG690

There's a lecture where he explicitly says this isn't true.

---

Said what specifically isn't true? What I incorrectly assumed or the lesson I concluded from it?

Sorry. The lesson you concluded from it.

Scored a 5.5/6 on lecture 4. But the deadline was yesterday so probably get 50% off. (100% off is for the HARD deadline) The deadline for lecture 5 is feb 17 so I'll have to do sometime this week.

Am I the only one having trouble reading the answers properly for problem set 7?

Well, I'll admit I haven't done the last two weeks because I couldn't be assed to.

I've had two emails talking about finals, so I assume it's crunch week.

And by the way, ignorant of my desire for any grade in a beta-online course for one university for which I didn't even use my real name, your last question would have been way better asked if you actually said what part of the answers of problem set 7 you were having trouble with and why.

Unless they're formatted unreadably or something.

My point about answers has always been that this thread would be best served as a nexus for everyone to touch base on the work they're doing and have done. That that work is almost always the same set of problems for the same class for which they ask you not to cheat is irrelevant. The value of doing it together is doing it together - outsourcing your thoughts to the thread and then insourcing them right back.

The thread has thus far only been hey I did 5 by 5 and is anyone else still keeping with it?

Quote:

---

Originally Posted by surviva316

Well crap, this sounds fun and awesome and I feel left out. I guess my options are to spread myself way too thin by signing up, to drop my Justice course on edX and take this one instead, or to wait until next semester and take the class with . . . . . Pascal *vomits all over myself*.

---

<3

Starts in 10 days

I like how the whole point of this thread to Rilla is to argue about how upset he is that the answers weren't posted here. Apparently, nobody has had trouble with problem sets since nobody has asked questions. We all must be smart motha fuckas.

You apparently haven't done the work and I really haven't had trouble. The only trouble I've had was reading the poorly formatted answers in problem set 7. I think we're the only two here actually using this thread here and obviously it's used best for asking for help. Luckily, this thread wasn't necessary. So be it.

Quote:

---

Originally Posted by BooG690

I like how the whole point of this thread to Rilla is to argue about how upset he is that the answers weren't posted here. Apparently, nobody has had trouble with problem sets since nobody has asked questions. We all must be smart motha fuckas.

You apparently haven't done the work and I really haven't had trouble. The only trouble I've had was reading the poorly formatted answers in problem set 7. I think we're the only two here actually using this thread here and obviously it's used best for asking for help. Luckily, this thread wasn't necessary. So be it.

---

Yeah, I was angry for no good reason. I just haven't cared about this class and haven't had reason to.

If you found the force of will to hammer through this course, more power to you.

I still stand by my position that the best use of this thread would have been for an open communication of our thoughts in this thread.

I just always thought that it wasn't going to be worth anything donating to if it wasn't about helping each other think through even the dumbest corners of game theory analysis.

And there's a wall when we're not allowed to share answers explicitly. A sort of code that every abides by to not break the answer key for anyone else. A code that shifts everyone's approach to communicating in the thread.

I just think we should have said f-u to Standford's honor code in benefit of our discussion of the course.

This class has been on my mind enough that I want to take a crack at finishing it, but I'm more interested in how we can approach classes in the future, together.

Well, our sample size here is kind of small. I really do believe you and I were the only ones participating in this thread. One of us fell behind and the other is a damn mathematical genius. Want to try again? I'm taking these classes next:

https://class.coursera.org/cgt-001/class/index
https://www.coursera.org/course/behavioralecon
https://www.coursera.org/course/ggp
https://www.coursera.org/course/gamification

So, boog, how has your understanding of game theory improved your life?

Taking this class has made me less bored. For the record, I took a game theory class in college. I just took this class in case they start making classes like these pre-reqs for other classes.

So, you understand game theory.

How do you use game theory?

I don't atm. I feel it'll help me in different classes though that may be useful in game playing (chess, poker, etc.).

Quote:

---

Originally Posted by BooG690

I don't atm. I feel it'll help me in different classes though that may be useful in game playing (chess, poker, etc.).

---

Yeah, game theory has given me a new 'vision' for all sorts of interactions that I have in the day to day.

I tried, like a lame-o, to tie it to Beyonce and her performance.

As if she had a decision between singing for herself and sounding like ass or lip-syncing and dancing it out - while 'sounding' amazing.

But the interesting thing I noticed was that she mixed responses in one response.

She lip-synced, then sang with her old gals au-naturale, then lip synced.

I was highly amused by this response that seemed to be outside the description of game theory.

The idea of utility is something I've always loved since my first economics class. That's probably my favorite sticking point when trying to use game theory or economics in my everyday life.

Also, that maximizing utility is the name of the game. Doing shit for spite is really stupid but it's amazing how many times that action is played.

Quote:

---

Originally Posted by BooG690

The idea of utility is something I've always loved since my first economics class. That's probably my favorite sticking point when trying to use game theory or economics in my everyday life.

Also, that maximizing utility is the name of the game. Doing shit for spite is really stupid but it's amazing how many times that action is played.

---

In the same spirit, I have always loved the idea of metrics.

Metrics as in 'to be measured'.

It has units. Speed is mph, m/s. Velocity is <x, y, z> m/s etc

And I've always wondered about the units of utility. They seem to be, most commonly, dollars. But dollars themselves are not a unit. Not like a gram or a meter, at least. They're more emotional.

So what is a utility to you?

Utility, to me, would be dollars or "happiness points." And sometimes you can use dollars to "buy happiness points."

Like, for instance, I have the choice to go to the amusement park or to the movies. The amusement park costs $20 while the movies cost $10. If I go to the amusement park, I go with a group of friends that I really enjoy. Also, I enjoy the amusement park more. The amusement park trip is worth 30 happiness points.

When I go to the movies, I go with one friend that is meh. Also, the movie will probably just be OK. The movies are worth 15 happiness points.

Let's make one dollar and a happiness point equal in terms of utility.

I lose 10 utility units to go to the movies but gain 15 for a net of 5.
I lose 20 utility units to go to the amusement park but gain 30 for a net of 10.
I obviously choose to go to the amusement park every time.

Obviously, this question was simplified for mathematics sake. Most people value money more than I do making my dollar to "happiness point" ratio pretty low. But yeah, there.

Quote:

---

Originally Posted by BooG690

Utility, to me, would be dollars or "happiness points." And sometimes you can use dollars to "buy happiness points."

Like, for instance, I have the choice to go to the amusement park or to the movies. The amusement park costs $20 while the movies cost $10. If I go to the amusement park, I go with a group of friends that I really enjoy. Also, I enjoy the amusement park more. The amusement park trip is worth 30 happiness points.

When I go to the movies, I go with one friend that is meh. Also, the movie will probably just be OK. The movies are worth 15 happiness points.

Let's make one dollar and a happiness point equal in terms of utility.

I lose 10 utility units to go to the movies but gain 15 for a net of 5.
I lose 20 utility units to go to the amusement park but gain 30 for a net of 10.
I obviously choose to go to the amusement park every time.

Obviously, this question was simplified for mathematics sake. Most people value money more than I do making my dollar to "happiness point" ratio pretty low. But yeah, there.

---

Is there a dollar~happiness ratio invariance? Is 10 dollars today worth .4 happiness points tomorrow?

This is the trouble I've always had with these descriptions. I kinda get where they're going, but with the exchange rates of happiness~dollars changing by the hour, I just don't see how it works.

Well, if I went to the amusement park yesterday, the happiness points I gain from going to the amusement park today is obviously quite low. I guess I try to think of the maximum number of dollars I'd pay to do XYZ. I usually set that as the happiness points I'm getting. If I pay lower than that, net gain of HPs. If I am asked to pay more, I ain't going.

Quote:

---

Originally Posted by BooG690

Well, our sample size here is kind of small. I really do believe you and I were the only ones participating in this thread. One of us fell behind and the other is a damn mathematical genius. Want to try again? I'm taking these classes next:

https://class.coursera.org/cgt-001/class/index
https://www.coursera.org/course/behavioralecon
https://www.coursera.org/course/ggp
https://www.coursera.org/course/gamification

---

Why are you taking these courses?

Quote:

---

Originally Posted by Pascal

<3

Starts in 10 days

---

GIVE ME THE GOD DAMNED READING LIST, EDX!

I google imaged "give me the reading list" just in the 1 in 10 trillion chance that some sort of hilarious meme existed for this, and this was my first hit:

http://1.bp.blogspot.com/-flvZt9GmQF...s_20110615.JPG

Quote:

---

Originally Posted by coursera

Gamification is the application of game elements and digital game design techniques to non-game problems, such as business and social impact challenges.

---

"Game elements and design techniques" seems really vague to me. Do they mean applying game theory? Constructing game simulations? Does it mean anything and everything it could possible mean?

It basically means gamifying things like fitness ala Fitocracy.

Quote:

---

Originally Posted by BooG690

It basically means gamifying things like fitness ala Fitocracy.

---

So a way to coax people to act?

Basically, yes.

Quote:

---

Originally Posted by a500lbgorilla

Why are you taking these courses?

---

Boredom? I enjoy playing games and understanding how the human mind works (so I can manipulate it).

Quote:

---

Originally Posted by BooG690

Boredom? I enjoy playing games and understanding how the human mind works (so I can manipulate it).

---

So it's not boredom :P

I'm all for the second leg of your reasoning.

Quote:

---

Originally Posted by BooG690

It basically means gamifying things like fitness ala Fitocracy.

---

So like, Sesame Street (by analogy) would be like televisonifying education?

Or the way Khan Academy has gamified learning maths:
https://www.khanacademy.org/exercisedashboard

I missed the first round, are these all free? I'd be interested in 1, 3 or 4 if 4hrs a week is likely sufficient (I have some relevant background)

These are all free. Start the first one. It's pretty interesting. It's only in the second week. Four hours a week is more than enough.